April 2022 - openstack-discuss - lists.openstack.org

Upgrading Openstack nodes
by Marc-Antoine Godde 06 Apr '22

06 Apr '22

Hello, We are running an Openstack cloud composed of 3 controller nodes and 4 compute nodes. Our deployment was realized with OpenStack-ansible and we are running OpenStack Ussuri on Ubuntu 18.04. Our plan is to upgrade nodes to Ubuntu 20.04, that way we would be able to update to OpenStack Victoria and further. We would like to withdraw each node from the cluster, reinstall a clean linux and redeploy the nodes. There is garbage remaining from previous upgrades. We figured out the way in the documentation to remove a compute node from the cluster with Openstack-ansible but we can’t find any related documentation for controller nodes. Any help would be very much appreciated. By the way, if you’d have any other suggestions on how to perform that upgrade, fell free to help. Best, Marc-Antoine Godde

2 2

[dev][infra][qa][tact-sig] Zuul behavior change with Depends-On across queues
by Jeremy Stanley 06 Apr '22

06 Apr '22

For those who haven't seen the more detailed announcement[*] about it, just a quick note that if you get a sudden -2 back from Zuul when approving a change with a Depends-On to a change in a different project which hasn't merged yet, that's likely an indication those projects don't share a dependent queue. It's not a bug, but an intentional clarification of Zuul's enqueuing behavior. For most other Zuul deployments (and even our other Zuul tenants in OpenDev) this is purely cosmetic, but since OpenStack's Zuul tenant is configured to require a positive Verified vote before enqueuing into the gate pipeline, it means some changes may end up unexpectedly needing another pass through check first. It's worth re-evaluating whether or not this "clean check" rule remains a useful requirement for gating. It was added some years ago because a number of gate breaking bugs were traced back to unstable changes being rechecked enough times that eventually they got lucky and were able to merge, and then their instability contributed to destabilizing the integrated gate as a whole. Similarly, changes were being approved without reviewers confirming their jobs were passing first, and this led to additional resource waste. There is a bit of discussion around "blind rechecks" at the PTG this week, and so this topic is related; it might be a good idea to consider it in conjunction with the greater recheck conversation. [*] http://lists.opendev.org/pipermail/service-announce/2022-April/000033.html -- Jeremy Stanley

2 2

[openstack-ansible][PTG] Session on 6th of April is cancelled
by Dmitriy Rabotyagov 06 Apr '22

06 Apr '22

1 0

[Oslo] IRC meeting.
by Daniel Mats Niklas Bengtsson 06 Apr '22

06 Apr '22

Hi there, I would like to know if you agree that we have the meeting once a week? Instead of the first and third Monday of the month. It will be easier to manage and if sometimes it is canceled it does not matter Even if the meetings each week are short, this way we will have regular follow-up.

2 1

Metadata service {devices list}
by Ahmed Abdelhamid 06 Apr '22

06 Apr '22

Hi All, I am running into a strange issue with the metadata service. Per metadata-service manual <https://docs.openstack.org/nova/latest/user/metadata.html>, the devices attached to a VM should be visible in $ curl http://169.254.169.254/openstack/2018-08-27/meta_data.json However, whenever i execute it from a VM , the devices array is empty and looks like this, any idea why ? Thanks { "random_seed": "yu5ZnkqF2CqnDZVAfZgarG...", "availability_zone": "nova", "keys": [ { "data": "ssh-rsa AAAAB3NzaC1y...== Generated by Nova\n", "type": "ssh", "name": "mykey" } ], "hostname": "test.novalocal", "launch_index": 0, "meta": { "priority": "low", "role": "webserver" }, "devices": [ ], "project_id": "f7ac731cc11f40efbc03a9f9e1d1d21f", "public_keys": { "mykey": "ssh-rsa AAAAB3NzaC1y...== Generated by Nova\n" }, "name": "test"}

3 10

wireguard - ? - puzzle
by lejeczek 06 Apr '22

06 Apr '22

Hi guys. Has anybody solved that puzzle? Or perhaps it's not a puzzle at all, I'd imagine might be trivial to experts. First I thought - and only thought so far thus asking here - 'allowed_address_pairs' I'd need but that obviously does not do anything as 'wireguard' creates its own ifaces. So.. how do you get your 'wireguard' in openstack to route (no NAT) to instances' local network(s)? many thanks, L.

3 3

[Octavia] Weekly meeting cancelled
by Gregory Thiemonge 05 Apr '22

05 Apr '22

Hi, As this is the PTG week, the Octavia weekly meeting is cancelled. Thanks, Gregory

1 0

[tripleo][RDO] Tags for Yoga GA version in tripleo projects
by Alfredo Moralejo Alonso 05 Apr '22

05 Apr '22

Hi, As we are doing builds for RDO Yoga release, I'd like to know what we should include in yoga release. In Xena, tags were created with the commits for the most recent promotion at around GA and shipped rpm packages for them. Additionally, we pinned RDO Trunk Xena to those versions. Will we get similar tags for Yoga? Best regards, Alfredo

2 3

Problems running Watcher on DevStack
by Christian Rohmann 05 Apr '22

05 Apr '22

Hello there, I am trying to setup devstack which is running Watcher. But I am running always into problems with setup not finishing and throwing errors (see below). To clarify: I am trying to setup the devstack with multiple nodes (three nodes to be exactly, one control/compute and two compute) following the instructions at https://docs.openstack.org/watcher/latest/contributor/devstack.html which leads to this configuration control / first compute node: > [[local|localrc]] > ADMIN_PASSWORD="<password>" > DATABASE_PASSWORD=$ADMIN_PASSWORD > RABBIT_PASSWORD=$ADMIN_PASSWORD > SERVICE_PASSWORD=$ADMIN_PASSWORD > SERVICE_HOST="<ip of the control node>" > HOST_IP="<ip of host>" > FIXED_RANGE="10.4.128.0/20" > FLOATING_RANGE="subnet range" > > DEVSTACK_RELEASE="stable/yoga" > > LOGFILE=/opt/stack/logs/stack.sh.log > LOGDAYS=7 > LOG_COLOR=False > > enable_plugin watcher https://opendev.org/openstack/watcher > $DEVSTACK_RELEASE > > enable_plugin watcher-dashboard > https://opendev.org/openstack/watcher-dashboard $DEVSTACK_RELEASE > > enable_plugin ceilometer https://opendev.org/openstack/ceilometer.git > $DEVSTACK_RELEASE > CEILOMETER_BACKEND=gnocchi > > enable_service ceilometer-api > enable_service ceilometer-acompute > > enable_plugin gnocchi https://github.com/gnocchixyz/gnocchi > > # OpenStack Telemetry (Ceilometer) Alarming > # enable_plugin aodh https://opendev.org/openstack/aodh $DEVSTACK_RELEASE > > # I did not use the panko project which is mentioned in the > documentation because it is deprecated > > [[post-config|$NOVA_CONF]] > [DEFAULT] > compute_monitors=cpu.virt_driver > [scheduler] > discover_hosts_in_cells_interval=2 additional compute node(s): > [[local|localrc]] > HOST_IP="<host ip>" > FIXED_RANGE=10.4.128.0/20 > FLOATING_RANGE="<floating range" > ADMIN_PASSWORD="<password>" > DATABASE_PASSWORD=$ADMIN_PASSWORD > RABBIT_PASSWORD=$ADMIN_PASSWORD > SERVICE_PASSWORD=$ADMIN_PASSWORD > DATABASE_TYPE=mysql > SERVICE_HOST="<ip of control node>" > MYSQL_HOST=$SERVICE_HOST > RABBIT_HOST=$SERVICE_HOST > GLANCE_HOSTPORT=$SERVICE_HOST:9292 > Q_HOST=$SERVICE_HOST > OVN_SB_REMOTE=tcp:$SERVICE_HOST:6642 > disable_all_services > ENABLED_SERVICES=n-cpu,placement-client,ovn-controller,q-ovn-metadata-agent > NOVA_VNC_ENABLED=True > NOVNCPROXY_URL="http://$SERVICE_HOST:6080/vnc_lite.html" > VNCSERVER_LISTEN=$HOST_IP > VNCSERVER_PROXYCLIENT_ADDRESS=$VNCSERVER_LISTEN > > DEVSTACK_RELEASE="stable/yoga" > > LOGFILE=/opt/stack/logs/stack.sh.log > LOGDAYS=7 > LOG_COLOR=False > > enable_plugin ceilometer https://opendev.org/openstack/ceilometer > $DEVSTACK_RELEASE > disable_service ceilometer-acentral > disable_service ceilometer-collector > disable_service ceilometer-api > > [[post-config|$NOVA_CONF]] > [DEFAULT] > compute_monitors=cpu.virt_driver In addition I have tried to use the configuration that can be found in the repository of the watcher-project at https://opendev.org/openstack/watcher/src/branch/master/devstack. The error is always the same with setup.sh ending in: > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:244 > : cp /opt/stack/ceilometer/etc/ceilometer/polling_all.yaml > /etc/ceilometer/polling.yaml > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:246 > : cp > /opt/stack/ceilometer/ceilometer/pipeline/data/event_definitions.yaml > /opt/stack/ceilometer/ceilometer/pipeline/data/event_pipeline.yaml > /opt/stack/ceilometer/ceilometer/pipeline/data/pipeline.yaml > /etc/ceilometer > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:248 > : '[' '' ']' > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:251 > : '[' False == True ']' > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:259 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > auth_type password > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:260 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > user_domain_id default > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:261 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > project_domain_id default > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:262 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > project_name service > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:263 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > username ceilometer > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:264 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > password <password> > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:265 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > region_name RegionOne > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:266 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > auth_url http://192.168.0.100/identity > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:268 > : configure_auth_token_middleware /etc/ceilometer/ceilometer.conf > ceilometer /var/cache/ceilometer > ++ lib/keystone:configure_auth_token_middleware:467 : echo 'WARNING: > configure_auth_token_middleware is deprecated, use > configure_keystone_authtoken_middleware instead' > WARNING: configure_auth_token_middleware is deprecated, use > configure_keystone_authtoken_middleware instead > ++ lib/keystone:configure_auth_token_middleware:468 : > configure_keystone_authtoken_middleware > /etc/ceilometer/ceilometer.conf ceilometer > ++ lib/keystone:configure_keystone_authtoken_middleware:447 : local > conf_file=/etc/ceilometer/ceilometer.conf > ++ lib/keystone:configure_keystone_authtoken_middleware:448 : local > admin_user=ceilometer > ++ lib/keystone:configure_keystone_authtoken_middleware:449 : local > section=keystone_authtoken > ++ lib/keystone:configure_keystone_authtoken_middleware:451 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken auth_type password > ++ lib/keystone:configure_keystone_authtoken_middleware:452 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken interface public > ++ lib/keystone:configure_keystone_authtoken_middleware:453 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken auth_url > http://192.168.0.100/identity > ++ lib/keystone:configure_keystone_authtoken_middleware:454 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken username ceilometer > ++ lib/keystone:configure_keystone_authtoken_middleware:455 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken password <password> > ++ lib/keystone:configure_keystone_authtoken_middleware:456 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken user_domain_name > Default > ++ lib/keystone:configure_keystone_authtoken_middleware:457 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken project_name service > ++ lib/keystone:configure_keystone_authtoken_middleware:458 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken project_domain_name > Default > ++ lib/keystone:configure_keystone_authtoken_middleware:460 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken cafile > /opt/stack/data/ca-bundle.pem > ++ lib/keystone:configure_keystone_authtoken_middleware:461 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken memcached_servers > localhost:11211 > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:270 > : [[ libvirt = \v\s\p\h\e\r\e ]] > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:277 > : _ceilometer_configure_storage_backend > ++ > /opt/stack/ceilometer/devstack/plugin.sh:_ceilometer_configure_storage_backend:208 > : '[' gnocchi = none ']' > ++ > /opt/stack/ceilometer/devstack/plugin.sh:_ceilometer_configure_storage_backend:210 > : '[' gnocchi = gnocchi ']' > ++ > /opt/stack/ceilometer/devstack/plugin.sh:_ceilometer_configure_storage_backend:211 > : sed -i > 's/gnocchi:\/\//gnocchi:\/\/?archive_policy=ceilometer-low\&filter_project=gnocchi_swift/' > /etc/ceilometer/event_pipeline.yaml /etc/ceilometer/pipeline.yaml > ++ > /opt/stack/ceilometer/devstack/plugin.sh:_ceilometer_configure_storage_backend:212 > : [[ ,watcher,watcher-dashboard,ceilometer,gnocchi =~ gnocchi ]] > + /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:1 : > exit_trap What am I missing? Any configuration parameter? Thanks in advance Christian

2 1

[TripleO] Move tripleo repos stable/ussuri to End Of Life OK?
by Marios Andreou 05 Apr '22

05 Apr '22

Hello TripleO o/ The tripleo-ci team proposes that we move the stable/ussuri branch for all tripleo repos [1] to End of Life [2]. The branch was moved to extended maintenance with [3] so we can already no longer make any new releases for ussuri/tripleo repos. Are there any objections or concerns about this? If so please speak up here or directly on the patch. I think that any patches posted to ussuri nowadays are mainly about cherrypicking back to train, rather than merging something to ussuri specifically. I have posted the proposal to move to EOL at [4] and have WF-1 to wait for this discussion. If there are no objections then once we move to EOL the tripleo-ci team will remove all Ussuri related CI, check/gate/promotions. regards, marios [1] https://releases.openstack.org/teams/tripleo.html#ussuri [2] https://docs.openstack.org/project-team-guide/stable-branches.html#maintena… [3] https://review.opendev.org/c/openstack/releases/+/817623 [4] https://review.opendev.org/c/openstack/releases/+/834049

2 5